Mat136 Integration
Mat136 Integration
Mat136 Integration
Li Chen
November 15, 2018
Contents
1 How to Use These Notes 1
3 Techniques of Integration 9
3.1 (5.2) Linearity [? ? ? ? ?] . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 (5.5) The Substitution Rule . . . . . . . . . . . . . . . . . . . . . 10
3.3 (7.1) Integration by Parts . . . . . . . . . . . . . . . . . . . . . . 14
3.4 (7.1) Trig Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4 Trig Substitution 19
5 Other Techniques 19
1
The majority of the contents in this note is geared to answer this questions.
In these notes, I will explicitly indicate which are the important materials.
Although one should be comfortable with all the materials in this note, but
certain critical concept must be emphasized to illuminate the essence of the
course. First, definitions are in bold. I will use a scale of stars to measure the
importance of the material. The importance of the concept decreases as the
number of star decreases. Whenever you see a 5-star tag [? ? ? ? ?], this means
you must know this inside out. And [?] is just a bit more important than the
regular text you see in the notes. Finally, I will italicize item that 1) you need
to be careful with as they might be confusing, or 2) to leave a comment.
The corresponding section numbers in [Stewart] are listed in bracket in the
heading of each section. And I will NOT write the integration constant +C in
the notes to save some ink, but you should always write it!!
2 f (x) = x2
0 x
−4 −3 −2 −1 0 1 2 3 4
−1
−2
−3
2
In this example, the shaded region represents the area under the curve y =
f (x) = x2 from x = −2 to x = 2. In general, to find the area under the curve
y = f (x) from x = a to x = b, we divide the interval [a, b] into segments
In fact the choice of sample point does not matter. In fact, we can choose any
sample point x∗i in the interval [xi−1 , xi ] and compute the sum
2.1.2 Displacement/Distance
A physically motivating example for the integral is the displacement traveled
by a car with velocity f (t) at time t. Suppose that from time t = a to t = b a
car travels at a velocity f (t). If f (t) = v is a constant. Then the displacement
traveled in ∆t units of time is simply
Now suppose that f (t) is variable. If it is continuous, then its value varies
slightly if t changes by a small amount. Hence we may approximate the distance
traveled by dividing the time interval [a, b] into small segments
3
units of length, where t∗i is any sample point in the interval [ti−1 , ii ]. The total
displacement, d, is, therefore, approximately,
Taking the segments [ti−1 , ti ] smaller for better approximation, we see that
and the displacement traveled in the time interval [a, b] by a car of velocity f (t)
is
n
X
d = lim f (t∗i )∆t [? ? ?] (2.1.13)
n→∞
i=1
provided this limit exists and is independent of the choice of sample points x∗i .
If the limist exists, we say that f is integrable.
The following names are given to the parts of the integral
Z b
f (x) dx
|{z} (2.2.3)
a |{z}
|{z} integrand integrate with respect to x
integral sign
4
We remark that the symbol dx only indicate that we are integrating with respect
to x, should there be any more variables appearing in the integrand.
The sum
n
X
f (x∗i )∆x (2.2.4)
i=1
Pn 2
3 n(n+1)
7. i=1 i = 2
5
2.3 (4.9) Antiderivatives
First, I want to dispel any possible confusion in notation. For our purpose, a
real function is a map from an interval I to the real line. I will use both f and
f (x) to mean such a function. However, confusion may arise when we want to
talk about a function evaluated at a point x = x0 . In this case the symbol f (x0 )
would mean the function f evaluated at the point x = x0 . In the former case,
the argument x in f (x) is generic, while in the latter case the point x0 is given
and fixed. In what follows, I will use f to denote a function. But the meaning
of f (x), either as a function or as the evaluation of f at the point x, depends
on the context. If needed, I will specify its meaning.
Given any function f (x) on [a, b], an antiderivatives is any function F (x)
on [a, b] such that
F 0 (x) = f (x). (2.3.1)
If F (x) is an antiderivative of f (x), so is F (x) + C for any constant C and these
are all the possible antiderivatives for f (x). For notation, we will use both
d
F (x), F 0 (x) (2.3.2)
dx
to mean the derivative of F (x).
d
We remark that the symbol dx F (x) may have ambiguous meaning. It could
d d
mean 1) dx (F (x)) – the derivative, with respect to x, of F (x) or 2) dx F (x) –
the evaluation of the
derivative of F at x. However, with either interpretation,
d d
dx (F (x)) = dx F (x). So there is in fact no ambiguity. To clarify notation
even more, the symbol
d
F (x) (a) = F 0 (a) (2.3.3)
dx
means the derivative, with respect to x, of F (or equivalently, the function F (x))
evaluated at the point x = a. However,
d
F (a) (x) = 0 (2.3.4)
dx
In conclusion, the order or taking derivative of a function and evaluating the
function at a point matters!
In Figure 1, I have listed for your a table of important derivatives. The left
column you should know off by heart [? ? ?].
6
Figure 1: Table of Antiderivatives [Stewart]
The way the theorem are written may not be best suited for our purpose.
What you should remember, for the purpose of this course and computing in-
tegrals, is the following formuli [? ? ? ? ?]
Z x Z b
d
f (t)dt = f (x), F 0 (t)dt = F (b) − F (a) . (2.4.3)
dx a a
What you should remember is that integrals and derivatives are almost inverses
of each other! We make a final note that the equation on the right in the box
is also called the net change theorem.
7
where F 0 (x) = f (x). Don’t forget your constant C! We also remark that the
Rb R
definite integral a f (x)dx is a number while the indefinite integral f (x)dx is
a function.
Figure 2 is table of indefinite integrals. You are expected to know all the
materials here, especially the first ten identities (all left column and the first
one from the right column).
8
3 Techniques of Integration
Always remember:
[? ? ? ? ?] The general philosophy is that techniques should give you a tool to
convert unknown integrand to known integrands.
As the sentence implies, this philosophy consists two parts
3. Integration by Parts,
4. Trigonometric Substitution, and
5. Partial Fraction.
The last two subsections are strategy for integration and (possibly non-standard
but useful) special techniques.
where c is a constant.
Here is an example. Integrate (1 + x2 )2 + 9ex + πx dx. We see that by
R
linearity
Z Z Z Z
2 2 x π 2 2 x 1
(1 + x ) + 9e + dx = (1 + x ) dx + 9 e dx + π dx. (3.1.3)
x x
9
Now we integrate term by term.
Z Z Z Z Z
(1 + x ) dx = 1 + 2x + x dx = 1dx + 2 x dx + x4 dx
2 2 2 4 2
(3.1.4)
2 1
= x + x3 + x5 , (3.1.5)
3 5
and e dx = ex , and x1 dx = ln x.
R x R
Just like the FTC, as stated, the rules are not as helpful as they are in
practice. In practice, the most important part of this technique is to identify the
substitution u = g(x). Below I included a (non-exhaustive) list of representative
samples of techniques.
Identify g(x). [? ? ? ? ?] This techniques is simple to explain and very
hard to use in practice: try to find f and g such that the integrand looks like
f (g(x))g 0 (x)dx where you know how to integrate f (with antiderivative F ).
R
Z
f (u)g 0 (x)dx. (3.2.3)
This tells us
Z Z
f (g(x))g 0 (x)dx = f (u)du (3.2.5)
10
Scaling. [? ? ? ? ?] Here is the idea:
1. Regard the integrand as f (ax) for some constant a such that you know
an antiderivative of f , say F .
2. Use the substitution u = ax with du = adx. Then
Z Z
1 1 1
f (ax)dx = f (u) du = F (u) = F (ax). (3.2.6)
a a a
R
Compute the integral 3 cos(9x)dx.
1. Use the substitution is u = 9x.
Equivalently,
1
dx = du. (3.2.8)
9
1
R
Compute the integral 1+x dx.
11
Substitute for the Bad Part. [? ? ?] Here is the idea:
1. Suppose that we can write the integrand as f1 (g(x))f2 (x) where f1 (g(x))
is hard to integrate explicitly or very ugly, but f1 (x) and f2 (x) are simple
and nice.
f2 (x)
4. If is a function of g(x), then we are in business.
g 0 (x)
√
Compute x3 x2 + 1dx.
R
√
1. The square root here is causing headache here. But f1 (x) = x is nice.
2. So we set u = x2 + 1.
3. We get g 0 (x) = 2x and du = 2xdx. And we get
√
Z p Z
x3 x2 + 1dx = x3 udx (3.2.14)
√
Z
1
= x2 udu (3.2.15)
2
√
Z
1
= (u − 1) udu (3.2.16)
2
Z
1
= u3/2 − u1/2 du (3.2.17)
2
where u = g(x).
12
3. We hope that the term f (g(x))(1 − g 0 (x)) is less complicated than f (g(x))
and can be integrated.
R 1
We illustrate this with and example. Compute the example 1+e x dx.
1 x 1
1. We regard 1+e x = f (e ) where f (x) = 1+x . We know that f has an-
tiderivative ln(1 + x). We set u = ex .
1
2. Subtract and add ex · 1+ex to get
1 + ex ex
Z Z Z
1
dx = dx − dx (3.2.21)
1 + ex 1 + ex 1 + ex
ex
Z Z
= 1 dx − dx (3.2.22)
1 + ex
Z
u 1
=x− du (3.2.23)
1+uu
Z
1
=x+ du (3.2.24)
1+u
2. Look for h(x) such that (h(x) · 1)0 = h(x) sec(x). To do this, first, we look
for functions such that when you differentiate, you pick up a factor of sec.
We observe that (tan(x))0 = sec2 (x) works. But certainly (tan(x))0 =
sec2 (x) 6= tan(x) sec(x). But luckily, tan(x) sec(x) = (sec(x))0 . So we
simply add them and note that
13
So we use the substitution u = tan(x) + sec(x) and conclude that
Z
sec(x)dx = ln(tan(x) + sec(x)) (3.2.28)
by equation (3.2.26).
Implicit Function [??] Here is the idea
1. Suppose that the integrand can be written as f (x)g(x)
2. Assume that g(x) has a (simple) antiderivative G(x) and there is a func-
tion h such that h(G(x)) = f (x) and h is easy to integrate. Very often
h(x) = f (G−1 (x)) if the function G(x) has an inverse. This condition is
guaranteed if g(x) > 0 for all x or g(x) < 0 for all x.
3. Use the substitution u = G(x). We see that du = G0 (x)dx = g(x)dx and
Z Z Z Z
f (x)g(x)dx = f (x)G0 (x)dx = h(G(x))G0 (x)dx = h(u)du
(3.2.29)
1√
R
We compute 2(x+3) x+2
dx.
1 √1
1. We have the integrand is the product of x+3 and 2 x+2
.
2. We can integrate either one. But we don’t want to be left with an ugly
1
square root in the denominator at the end. So we pick g(x) = √x+2 . Note
that g(x) > 0 for all x in the proper domain so its antiderivative
√ has an
inverse. We see that an antiderivative of g is G(x) = x + 2.
√ 1
3. So we substitute u = x + 2 and we see that x = u2 −2 and du = 2√x+2 dx
and
Z Z Z
1 1 1
√ dx = 2
du = 2
du = tan−1 (u)
2(x + 3) x + 2 (u − 2 + 3) u +1
(3.2.30)
14
1. The integrand looks like a product of two function f (x)g 0 (x), one of them
g 0 (x) you know how to integrate and the other f (x) you know how to
differentiate.
2. RYou that if you can shift one
R derivative from g to f , then you can integrate
f 0 (x)g(x)dx instead of f (x)g 0 (x)dx.
3. Then we use formula (3.3.1). In practice, we use the pattern
u= dv = (3.3.2)
du = v= (3.3.3)
Then the rule is written as
Z Z
udv = uv − vdu. (3.3.4)
tan−1 (x)dx.
R
Here is an example. Integrate
1. Set u = tan−1 (x) and dv = 1dx. We get
u = tan−1 (x) dv = dx (3.3.11)
1
du = v = x. (3.3.12)
1 + x2
15
2. The integral becomes
Z Z
−1 −1 x
tan (x)dx = x tan (x) − dx (3.3.13)
1 + x2
The second term on the RHS can be integrated using the substitution
u = x2 .
Reduction of Power: Polynomials [? ? ? ? ?] Here is the idea:
1. The integrand looks like xn f (x) for some f (x) that you know how to
integrate (as many times as possible). We want to reduce the power of xn
by as much as we can since we know how to integrate f (x) repeatedly.
2. Use u = xn and dv = f (x)dx.
3. We get
u = xn dv = f (x)dx (3.3.14)
n−1
du = nx v = F (x) (3.3.15)
x5 ex dx.
R
Here is an example:
1. We can integrate ex for as many times as we need.
2. Use u = x5 and dv = ex dx.
3. We get
u = x5 dv = ex dx (3.3.17)
4 x
du = 5x v=e (3.3.18)
Using the same idea, one can repeat integration by part to integrate, by
getting rid of all powers of x.
Reduction of Power: General Setting [? ? ? ? ?] Here is the idea:
1. The integrand looks like (f (x))n g 0 (x). You can integrate g 0 (x) and you
want to reduce the power n.
16
2. Use u = (f (x))n and dv = g 0 (x)dx.
Here is an example: (ln x)4 dx.
R
1. It’s not so obvious how to choose factors. But we can write the integrand
as 1 · (ln x)4 .
2. Use u = (ln x)4 and dv = dx.
3. We get
u = (ln x)5 dv = dx (3.3.20)
1
du = 5(ln x)4 v = xdx (3.3.21)
x
4. The integral becomes
Z Z
(ln x)5 dx = x(ln x)5 − 5 (ln x)4 dx (3.3.22)
and we repeat.
Recursive Integration This one is best illustrated by example. The main
idea is that integration by part does not make the integrand look simpler, how-
ever after some iterations of integration by parts, you get back to where you
started, but with
R a different coefficient!
Compute ex cos(x)dx. We use
u = cos(x) dv = ex dx (3.3.23)
du = − sin(x)dx v = ex (3.3.24)
to get
Z Z
x x
e cos(x)dx = cos(x)e + sin(x)ex dx. (3.3.25)
This is truly usely, as the integrand does not become any simpler. But let’s not
give up. We do integration by parts again. This time we use
u = sin(x) dv = ex dx (3.3.26)
du = cos(x)dx v = ex (3.3.27)
to get
Z Z
x x
e cos(x)dx = cos(x)e + sin(x)ex dx (3.3.28)
Z
= cos(x)ex + sin(x)ex − cos(x)ex . (3.3.29)
17
3.4 (7.1) Trig Integrals
In order for us to compute trig integrals. We need 2 things
1. Integral of simple trig functions. You need to memorize all the integrals
from Fig. 2.
2. Basic trig identities. I listed them in Fig. 3. You need to know [? ? ?]:
1) Pythagorean identities, 2) double angle, 3) half angle, and 4) product
formuli. It is useful to know the addition and subtraction formuli.
Trig integrals that we concern ourselves with are of the follow three forms:
Product of sinn (x) and cosm (x) We give a summary of the strategy for
computing this kind of integrals in Fig. 4.
Product of tann and secm . We give a summary of the strategy for com-
puting this kind of integrals in Fig. 5.
18
Product of sin(ax) and cos(bx). We give a summary of the strategy for
computing this kind of integrals in Fig. 6.
4 Trig Substitution
5 Other Techniques
p
I + J and I − J try tan(x).
References
[Stewart] J. Stewart, Single Variable Calculus Early Transcendentals, 8th edi-
tion. ISBN-13: 978-1-305-27033-6
19
Figure 4: Product of sinn (x) and cosm (x), [Stewart]
20
Figure 5: Product of tanm (x) and secm (x), [Stewart]
21